A CT Image Denoising Approach Using Wiener Filter and NSST Domain

Denoising (LDCT) low dose computed-tomographical imaging is essential for improving diagnostic preciseness and image clarity, especially in the presence of gaussian noise. In LDCT imaging, gaussian noise adversely affects the image quality. To address this concern, the proposed a novel method combines the (NSST) no subsampled shearlet transform with wiener filtering and bivariate shrinkage, resulting in improved denoising outcomes. Initially, the NSST approach, decomposes the CT image into approximation and detail parts. In the approximation part, the wiener filter was applied, and bivariate shrinkage thresholding was applied in the detail part to get denoised coefficients. The inverse shearlet transform is employed for obtaining the final-denoised image. The experimental outcome of the proposed method is enumerated relating to the (PSNR) peak signal-to-noise ratio, (SNR) signal to noise ratio, and (UIQI) universal image quality index of particular CT images. The PSNR, SNR, and UIQI values at different noise variances (σ = 10, 15, 20, and 25) have been taken to analyze the performance of quantitative metrics with standard denoising filters. Here, the proposed method provides better results with regards to noise suppression and visual quality of CT images.